Problem

WorldServer does pre-fill all target segments with tags and whitespace from the source, if no pre-translation is available in WorldServer.

Since translate5 does pre-translate only empty segments, for those segments there will be no possibility to pre-translate them.

Solution

We need a functionality, that deletes all content from target segments, that only contain whitespace and/or tags before the sdlxliff file is imported into the segment table.

If segment pair looks like this, the segment is already skipped by the import (at least should be like this):

<trans-unit><source>only tags and/or whitespace</source><target>no matter what is in here</target></trans-unit>

If segment pair looks like this, our new functionality should change nothing - the segment should be imported:

<trans-unit><source>some content, that contains text (and possibly tags and/or whitespace)</source><target>some content, that contains text (and possibly tags and/or whitespace)</target></trans-unit>

If segment pair looks like this, our new functionality should delete the content of the target:

<trans-unit><source>some content, that contains text (and possibly tags and/or whitespace)</source><target>only tags and whitespace </target></trans-unit>

This can / should be developed inside the sdlxliff parser of translate5.

If the functionality is active should be defined by a config, that is overwritable on client and import level.

It needs to be made sure, that this also works, when the connectWorldserver plug-in imports sdlxliff files into translate5.